In this step, we will evaluate which workloads we want to migrate. After identifying the candidates for migration, we should always run an inventory of our VMs or physical servers and calculate what the total cost of ownership (TCO) of maintaining the infrastructure is. Things such as hardware cost, support maintenance contracts, electricity bills, and even space rental come into play here. This will help us to understand how much we will be saving in an eventual migration to the cloud. This data is crucial to convince management and any C-level decision makers that may have any doubts about the benefits in cost of migrating infrastructure to a cloud provider.
The ideal scenario to begin a migration is to look for smaller applications that don’t require an entire infrastructure to be migrated in order for them to be put into production. Applications with few dependencies are perfect to begin your assessment. Dependencies such as which servers we need to migrate together, and network requirements for our application such as ports and IP operational ranges are to be taken into consideration. Questions such as the following will help us to prepare for a successful migration:
- Is my Linux distribution endorsed by the cloud provider I am migrating to?
- Am I running the Kernel version that's supported by my cloud provider?
- Do I have to install any additional kernel modules?
- Does my cloud provider require any type of agent running on my OS?
With these questions answered, we can start performing the actual migration.